National Cholesterol Awareness Month Article Clinical and Population Studies

Transcription

1 National Cholesterol Awareness Month Article Clinical and Population Studies A Common LPA Null Allele Associates With Lower Lipoprotein(a) Levels and Coronary Artery Disease Risk Theodosios Kyriakou, Udo Seedorf, Anuj Goel, Jemma C. Hopewell, Robert Clarke, Hugh Watkins, Martin Farrall, on behalf of the PROCARDIS Consortium Objective Increased levels of lipoprotein(a) are a highly heritable risk factor for coronary artery disease (CAD). The genetic determinants of lipoprotein(a) levels are mainly because of genetic variation in the apolipoprotein(a) gene (LPA). We have tested the association of a relatively common null allele of LPA with lipoprotein(a) levels and CAD risk in a large case control cohort. We have also examined how null allele genotyping complements apolipoprotein(a) isoform typing to refine the relationship between LPA isoform size and circulating lipoprotein(a) levels. Approach and Results The LPA null allele (rs ) was genotyped in the PROCARDIS (Precocious Coronary Artery Disease) case control cohort (4073 CAD cases and 4225 controls). Lipoprotein(a) levels were measured in 909 CAD cases and 922 controls; apolipoprotein(a) isoform size was estimated using sodium dodecyl sulfate agarose gel electrophoresis and a high-throughput quantitative polymerase chain reaction based method. Null carriers are common (null allele frequency, 3%) and have significantly lower circulating lipoprotein(a) levels (P= ) and reduced CAD risk (odds ratio, 0.79 [ ]; P=0.023) compared with noncarriers. An additive allelic model of apolipoprotein(a) isoform size, refined by null allele genotype and quantitative polymerase chain reaction values, showed a sigmoid relationship with lipoprotein(a) levels, with baseline levels for longer isoform alleles and progressively higher levels of lipoprotein(a) for shorter isoform alleles. Conclusions The LPA null allele (rs ) is associated with decreased circulating lipoprotein(a) levels and decreased CAD risk. Incorporating rs refined apolipoprotein(a) isoform size typing obtained by immunoblotting and quantitative polymerase chain reaction. A joint genomic and isoform analysis revealed details of the relationship between apolipoprotein(a) isoform size and circulating lipoprotein(a) level consistent with a threshold effect on lipoprotein secretion. (Arterioscler Thromb Vasc Biol. 2014;34: ) Lipoprotein(a) is a complex lipoprotein particle found in the circulation of higher primates. It comprises a lowdensity lipoprotein like particle with the addition of an apolipoprotein(a) molecule covalently bonded to apolipoprotein B. 1 In humans, increased lipoprotein(a) levels in the blood have been associated with increased risk of cardiovascular disease. 2 5 Lipoprotein(a) levels vary substantially among individuals within a population, are uncorrelated with classic risk factors, and remain relatively constant throughout an individual s lifespan. 6 A large proportion of lipoprotein(a) level variation is inherited and is mainly attributed to genetic variation within the LPA gene that encodes apolipoprotein(a), the characteristic protein moiety of the particle. 7 Genetic variation in the LPA gene is defined by a repetitive genomic region harboring 2 exons encoding one of the kringle domains (kringle IV [KIV]-2) in apolipoprotein(a), resulting in a highly polymorphic protein Key Words: coronary artery disease lipoprotein(a) that includes multiple copies of the kringle domain ranging in number from 3 to >40. 8 This variation has been proposed to directly affect circulating lipoprotein(a) levels by influencing the processing and secretion of apolipoprotein(a) isoforms from producing hepatocytes 9,10 The copy number of KIV-2 repeats is inversely related to levels with larger molecules, hindering efficient secretion and resulting in lower circulating lipoprotein(a) levels. 5 Depending on the population studied, a significant proportion of the genetic variation attributed to the LPA gene is independent of the copy number variation (CNV). 11 This is consistent with the observation that isoforms of the same size differ widely in concentration, suggesting that variation in the number of kringle domains is not the only important LPA contributor to lipoprotein(a) quantitative variation. There are reports of single nucleotide polymorphisms (SNPs) associated with lipoprotein(a) levels, which either serve as proxy Received on: February 13, 2014; final version accepted on: June 1, From the Division of Cardiovascular Medicine, Radcliffe Department of Medicine (T.K., A,G., H.W., M.F.), Wellcome Trust Centre for Human Genetics (T.K., A.G., H.W., M.F.), and Clinical Trial Service Unit and Epidemiological Studies Unit (J.C.H., R.C.), University of Oxford, Oxford, United Kingdom; and Leibniz-Institut für Arterioskleroseforschung an der Universität Münster, Münster, Germany (U.S.). The online-only Data Supplement is available with this article at Correspondence to Martin Farrall, FRCPath, Wellcome Trust Centre for Human Genetics, Old Rd Campus, Roosevelt Dr, Oxford OX3 7BN, United Kingdom. martin.farrall@cardiov.ox.ac.uk 2014 American Heart Association, Inc. Arterioscler Thromb Vasc Biol is available at DOI: /ATVBAHA

2 2096 Arterioscler Thromb Vasc Biol September 2014 Nonstandard Abbreviations and Acronyms CAD CNV KIV qpcr RQ SNP coronary artery disease copy number variation kringle IV quantitative polymerase chain reaction relative quantity single nucleotide polymorphism (tagging) markers for the CNV 5 or have no proven direct functional role or tagging association, such as the pentanucleotide repeat in the LPA promoter. 12 SNP rs is a functional polymorphism that results in alternative splicing and a prematurely terminated protein molecule that lacks the domain that is normally covalently linked to apolipoprotein B and is unable to form a mature lipoprotein(a) particle. 13 The functional effect of this relative common SNP, commonly referred to as a null allele, on circulating lipoprotein(a) levels has been found to be significant. 13 However, this SNP has not been included in any of the widely used commercial high-throughput SNP arrays and thus has not been examined in the large-scale genetic epidemiological studies that have highlighted the role of lipoprotein(a) as an independent risk factor for coronary artery disease (CAD), 5,14 16 ischemic stroke, 17 and other atherosclerotic diseases. 18 The main aim of the present study was to evaluate the effect of the LPA null allele on circulating lipoprotein(a) levels and CAD risk in PROCARDIS (Precocious Coronary Artery Disease), a European CAD case control cohort. A second aim was to refine the relationship between apolipoprotein(a)/lpa isoform size/cnv copy number and lipoprotein(a) levels by combining null allele genotyping with kringle isoform typing based on immunoblotting and genomic methods. Material and Methods Materials and Methods are available in the online-only Supplement. Figure 1. Effect of null allele on circulating lipoprotein (a) [Lp(a)] levels. These box-and-whisker plots include a central box that shows the median and upper and lower quartile Lp(a) levels; extreme values are indicated by open circles, and the upper and lower adjacent values are indicated by whiskers. Table. Summary Statistics of Carriers and Noncarriers of rs Carriers Noncarriers Demographics n Age* 62 (56 68) 63 (57 68) Sex (% males) 69% 68% Statin use, % Biochemistry Lp(a), mg/dl* 7.5 ( ) 12.3 ( ) HDL-C, mmol/l* 1.3 ( ) 1.2 ( ) LDL-C, mmol/l* 2.9 ( ) 2.9 ( ) KIV-2 repeats, n (%) Single band 93 (100) 1073 (61) Double band 0 (0) 691 (39) KIV-2 sum (immunoblots)* 28 (24 33) 56 (48 62) KIV-2 sum (qpcr)* 0.99 ( ) 1.03 ( ) HDL-C indicates high-density lipoprotein cholesterol; Lp(a), lipoprotein(a); LDL-C, low-density lipoprotein cholesterol; KIV, kringle IV; and qpcr, quantitative polymerase chain reaction. *Median (25% 75%). Results LPA Null Allele in the PROCARDIS Study: Effect on Circulating Lipoprotein(a) Levels and Association With CAD Risk. In the PROCARDIS cohort, minor allele frequency of the null rs SNP was 0.03, and genotype frequencies were within Hardy Weinberg equilibrium (χ 2 =0.58; 1 df; P=0.44). Lipoprotein(a) levels were measured in 2000 individuals with a mean concentration of 21.6 mg/dl (SD, 24.6; median, 11.5 mg/dl; interquartile range, mg/dl). After excluding individuals with missing genotypes, data were available for 1831 individuals (909 cases and 922 controls). The distribution of lipoprotein(a) levels in LPA null carriers and noncarriers is shown in Figure 1 and the Table. There was a significant association between circulating lipoprotein(a) levels and LPA null carrier status (P= ) with a 5.25 mg/dl (95% confidence interval [CI], ) reduction in median lipoprotein(a) levels in carriers. This association was also evident (P= ) in a joint analysis with 2 SNPs rs and rs that have been previously shown to be strongly associated with lipoprotein(a) levels 5 ; the LPA null allele was in linkage equilibrium (r2<0.002) with both these SNPs. We then tested the variant for association with CAD risk in 8211 individuals (4022 cases and 4189 controls) and observed a significant protective effect with 20% reduction in CAD risk (odds ratio, 0.79; 95% CI, ; P=0.023) in carriers. Correlation of Kringle Isoform Typing Using qpcr and Immunoblotting Figure 2 shows the relationship between proteomic (isoform size) and genomic (relative quantity [RQ]) measurements of kringle copy number. Carriers and noncarriers of the LPA null allele show positive correlations of similar magnitude

3 Kyriakou et al Lp(a) Null Allele and CAD 2097 Figure 2. The relationship between isoform size as identified by immunoblotting and quantitative polymerase chain reaction. Null allele carriers are represented by triangles, whereas noncarriers are represented by filled circles. RQ indicates relative quantity. (ρ=0.37, 95%CI, ; ρ=0.44, 95%CI, , respectively). The correlation between isoform size as measured by the 2 assay types when the null allele is not taken into account although positive is imperfect because of the inherent inaccuracies of the assay methods. This was not surprising and consistent with previous reports in smaller cohorts. 19 Regression Analysis of Lipoprotein(a) Levels Figure 3 summarizes a regression analysis of median lipoprotein(a) levels on RQ levels, which shows a nonlinear association pattern with a pseudo-r 2 value of Samples with RQ values >5th vigintile ( 25th percentile point) have similar lipoprotein(a) levels; samples with smaller RQ values show increasing lipoprotein(a) levels. The LPA null variant was a significant independent predictor of lipoprotein(a) levels (P= ), with carriers showing a 6.15 mg/dl (95% CI, ) reduction in levels. Figure 4 summarizes a regression analysis of median lipoprotein(a) levels on kringle isoform size; kringle alleles Figure 3. Median regression analysis of lipoprotein(a) [Lp(a)] levels on relative quantity (RQ) values. Ranked RQ values have been grouped into 20 equal-sized groups. Median Lp(a) levels are indicated by square markers, and upper and lower 95% confidence limits are indicated by whiskers. Figure 4. Median regression analysis of lipoprotein(a) [Lp(a)] levels on kringle isoform alleles. Median Lp(a) levels are indicated by square markers, and upper and lower 95% confidence limits are indicated by whiskers. were assumed to have an additive effect on lipoprotein(a) levels (ie, a codominant inheritance model), and the model included information on the null LPA allele and RQ values. Lipoprotein(a) levels show a nonlinear pattern of association with kringle alleles with a pseudo-r 2 value of 0.29; alleles with 23 repeats had similar lipoprotein(a) levels ( 6 mg/ dl per allele), whereas shorter alleles show progressively increasing lipoprotein(a) levels that plateau at 42 mg/dl. Covariation by RQ values was a significant independent predictor of lipoprotein(a) levels (P=0.0002), but RQ values have relatively little influence on the association pattern in comparison with the isoform data (Figure I in the online-only Data Supplement). The wide CIs for the quantitative polymerase chain reaction (qpcr) data (Figure 3) demonstrate the relative imprecision of this method compared with the isoform immunoblotting method (Figure 4). A supplementary regression analysis that disregards the null LPA allele shows a similar sigmoid pattern but with broader CIs (Figure II in the onlineonly Data Supplement). Discussion Genome-wide and gene-centric association studies have led to the identification of 48 genetic loci associated with CAD risk. The LPA locus on chromosome 6q shows the strongest genetic association with CAD discovered to date. 20,21 In this study, we have confirmed the effect of a common LPA null allele on circulating lipoprotein(a) levels and shown for the first time a clear association with decreased CAD risk. We and others have previously mapped quantitative trait loci for lipoprotein(a) and estimated that the vast majority (74 to >90%) of the total variation in circulating levels was specific to the LPA locus. 7,22,23 Several studies sought to identify genetic variation in LPA that explains variation in lipoprotein(a) levels and further define the association with CAD risk beyond the contribution of the well-established copy number variation in the KIV-2 domains. This resulted mainly in identification of SNPs that act as proxy markers of the KIV-2 CNV 5,24 or have no proven functional role. 12,14,25

4 2098 Arterioscler Thromb Vasc Biol September 2014 The rs single nucleotide polymorphism resulting in an LPA null allele was previously reported by Ogorelkova et al. 13 The minor allele of this SNP causes disruption of a donor splice site and results in a truncated allelic form of apolipoprotein(a), including only kringle domains 1 to 7 (KIV 1 7). Because of the presumed functional importance of specific kringle domains, mutations and truncations related to these are expected to lead to loss of function. Expression of the alternative spliced transcript from the null allele in HepG2 cells showed that the truncated apolipoprotein(a) molecule is expressed and secreted, but because it lacks the KIV-9 domain it is not possible to covalently link to apolipoprotein B-100 and form a stable lipoprotein(a) particle. 13 This truncated apolipoprotein(a) form was found to be extensively degraded in the plasma of carriers and was not detectable by immunoblotting. 13 A null allele effect was clearly evident in the samples of the PROCARDIS study where we found a clear association between circulating levels and carrier status of rs with a strong effect on lipoprotein(a) levels per null allele, resulting in a significant reduction of median levels. The wellestablished association of circulating levels and CAD risk prompted us to investigate the effect of the LPA null allele on CAD risk in the entire PROCARDIS collection. As expected, reduced levels of lipoprotein(a) because of the null allele had a significant protective effect against risk of disease. This finding highlights the contribution and importance of the null allele as the only known functional polymorphism other than the CNV in modulating lipoprotein(a) levels and CAD disease risk. It is important to note that the rs SNP resulting in the LPA null allele is not included in any of the available genome-wide, gene-centric, or exome genotyping arrays. Furthermore, these arrays do not include proxy SNPs in strong linkage disequilibrium (r2<0.6), and rs is not present in the HapMap2 or HapMap3 imputation training sets that have been widely used to drive genome-wide association studies. The findings presented here emphasize the importance of including the null allele in genotyping strategies to be used in future genetic and epidemiological studies. It was apparent that null allele carrier status could also supplement and refine genotyping data of lipoprotein(a) isoform size because of the KIV-2 CNV. Agarose gel separation followed by immunoblotting and qpcr are 2 methods routinely used in the past for genotyping the LPA CNV in studies including large number of samples. Lipoprotein(a) kringle isoform typing using immunoblotting is an expensive and low-throughput method that has limited accuracy. It has, however, the advantage that it can provide protein isoform size information encoded separately by each LPA allele (provided they are secreted at detectable levels). Because of the extensive apolipoprotein(a) gene size polymorphism, homozygotes are expected to be unusual in populations in Hardy Weinberg equilibrium; a heterozygosity index of 94% was reported among Europeans. 26 In the PROCARDIS study, a double band was observed in only a third of the samples assayed using immunoblotting. Null allele carrier status accounted for 12% of the single band immunoblots, with the others probably explained by nonsecreted large isoforms or low assay sensitivity. A qpcr method using genomic DNA as template has been recently used to overcome the inherent restrictions of immunoblotting and estimate the number of kringle repeats in samples from large epidemiological cohorts. 3,16 However, the high-throughput merits of speed and low cost of the qpcr method are tempered by the inability to distinguish the relative contributions of each allele because it can be used only for detection of the sum of the numbers of kringle domains. Furthermore, this method cannot identify alleles that do not produce a detectable apolipoprotein(a) molecule because of the steric impediment in secretion from producing cells. Apolipoprotein(a) isoform information derived from immunoblotting and qpcr was supplemented by null allele carrier status to perform a regression analysis of circulating lipoprotein(a) levels in the PROCARDIS samples. Isoform alleles showed a sigmoid pattern of association with lipoprotein(a) levels, suggesting that there is a threshold effect such that relatively short alleles with 21 KIV-2 repeats are readily secreted by hepatocytes. This in vivo result is consistent with and provides detail to the inferences drawn from an in vitro study contrasting the expression of 11 and 22 KIV-2 repeat containing plasmids in HepG2 cells. 9 In conclusion, the LPA null allele has a consistent and substantial effect in reducing circulating lipoprotein(a) levels and reducing CAD risk. Genotyping of the null allele enhances the data obtained by either immunoblotting or qpcr for measuring kringle KIV-2 copy number and significantly improves the characterization of LPA-lipoprotein(a) for quantitative genetic analyses. Acknowledgments We gratefully acknowledge technical assistance from Bertram Tambyrajah at Leibniz-Institut fur Arterioskleroseforschung an der Universitat Munster, Munster, Germany. Sources of Funding The PROCARDIS study was supported by the European Community Sixth Framework Program (LSHM-CT ), AstraZeneca, the British Heart Foundation, the Oxford British Heart Foundation Centre of Research Excellence, the Wellcome Trust (075491/Z/04), the Swedish Research Council, the Knut and Alice Wallenberg Foundation, the Swedish Heart-Lung Foundation, the Torsten and Ragnar Söderberg Foundation, the Strategic Cardiovascular Program of Karolinska Institutet and Stockholm County Council, and the Foundation for Strategic Research and the Stockholm County Council (560283). None. Disclosures References 1. Utermann G, Weber W. Protein composition of Lp(a) lipoprotein from human plasma. FEBS Lett. 1983;154: Erqou S, Kaptoge S, Perry PL, Di Angelantonio E, Thompson A, White IR, Marcovina SM, Collins R, Thompson SG, Danesh J. Lipoprotein(a) concentration and the risk of coronary heart disease, stroke, and nonvascular mortality. JAMA : Kamstrup PR, Tybjaerg-Hansen A, Steffensen R, Nordestgaard BG. Genetically elevated lipoprotein(a) and increased risk of myocardial infarction. JAMA. 2009;301:

5 Kyriakou et al Lp(a) Null Allele and CAD Virani SS, Brautbar A, Davis BC, Nambi V, Hoogeveen RC, Sharrett AR, Coresh J, Mosley TH, Morrisett JD, Catellier DJ, Folsom AR, Boerwinkle E, Ballantyne CM. Associations between lipoprotein(a) levels and cardiovascular outcomes in black and white subjects: the Atherosclerosis Risk in Communities (ARIC) Study. Circulation. 2012;125: Clarke R, Peden JF, Hopewell JC, et al.; PROCARDIS Consortium. Genetic variants associated with Lp(a) lipoprotein level and coronary disease. N Engl J Med. 2009;361: Dubé JB, Boffa MB, Hegele RA, Koschinsky ML. Lipoprotein(a): more interesting than ever after 50 years. Curr Opin Lipidol. 2012;23: Barlera S, Specchia C, Farrall M, Chiodini BD, Franzosi MG, Rust S, Green F, Nicolis EB, Peden J, Assmann G, Collins R, Hamsten A, Tognoni G, Watkins H; PROCARDIS Consortium. Multiple QTL influence the serum Lp(a) concentration: a genome-wide linkage screen in the PROCARDIS study. Eur J Hum Genet. 2007;15: Lackner C, Cohen JC, Hobbs HH. Molecular definition of the extreme size polymorphism in apolipoprotein(a). Hum Mol Genet. 1993;2: Brunner C, Lobentanz EM, Pethö-Schramm A, Ernst A, Kang C, Dieplinger H, Müller HJ, Utermann G. The number of identical kringle IV repeats in apolipoprotein(a) affects its processing and secretion by HepG2 cells. J Biol Chem. 1996;271: Hopewell JC, Seedorf U, Farrall M, Parish S, Kyriakou T, Goel A, Hamsten A, Collins R, Watkins H, Clarke R. Impact of lipoprotein(a) levels and apolipoprotein(a) isoform size on risk of coronary heart disease. J Intern Med doi: /joim Utermann G. Genetic architecture and evolution of the lipoprotein(a) trait. Curr Opin Lipidol. 1999;10: Valenti K, Aveynier E, Leauté S, Laporte F, Hadjian AJ. Contribution of apolipoprotein(a) size, pentanucleotide TTTTA repeat and C/T(+93) polymorphisms of the apo(a) gene to regulation of lipoprotein(a) plasma levels in a population of young European Caucasians. Atherosclerosis. 1999;147: Ogorelkova M, Gruber A, Utermann G. Molecular basis of congenital lp(a) deficiency: a frequent apo(a) null mutation in caucasians. Hum Mol Genet. 1999;8: Ober C, Nord AS, Thompson EE, et al. Genome-wide association study of plasma lipoprotein(a) levels identifies multiple genes on chromosome 6q. J Lipid Res. 2009;50: Trégouët DA, König IR, Erdmann J, et al.; Wellcome Trust Case Control Consortium; Cardiogenics Consortium. Genome-wide haplotype Significance association study identifies the SLC22A3-LPAL2-LPA gene cluster as a risk locus for coronary artery disease. Nat Genet. 2009;41: Lanktree MB, Anand SS, Yusuf S, Hegele RA; SHARE Investigators. Comprehensive analysis of genomic variation in the LPA locus and its relationship to plasma lipoprotein(a) in South Asians, Chinese, and European Caucasians. Circ Cardiovasc Genet. 2010;3: Helgadottir A, Gretarsdottir S, Thorleifsson G, et al. Apolipoprotein(a) genetic sequence variants associated with systemic atherosclerosis and coronary atherosclerotic burden but not with venous thromboembolism. J Am Coll Cardiol. 2012;60: Thanassoulis G, Campbell CY, Owens DS, et al.; CHARGE Extracoronary Calcium Working Group. Genetic associations with valvular calcification and aortic stenosis. N Engl J Med. 2013;368: Lanktree MB, Rajakumar C, Brunt JH, Koschinsky ML, Connelly PW, Hegele RA. Determination of lipoprotein(a) kringle repeat number from genomic DNA: copy number variation genotyping using qpcr. J Lipid Res. 2009;50: Peden JF, Farrall M. Thirty-five common variants for coronary artery disease: the fruits of much collaborative labour. Hum Mol Genet. 2011;20(R2):R198 R Deloukas P, Kanoni S, Willenborg C, et al. Large-scale association analysis identifies new risk loci for coronary artery disease. Nature Genetics, : Lamon-Fava S, Jimenez D, Christian JC, Fabsitz RR, Reed T, Carmelli D, Castelli WP, Ordovas JM, Wilson PW, Schaefer EJ. The NHLBI Twin Study: heritability of apolipoprotein A-I, B, and low density lipoprotein subclasses and concordance for lipoprotein(a). Atherosclerosis. 1991;91: Austin MA, Sandholzer C, Selby JV, Newman B, Krauss RM, Utermann G. Lipoprotein(a) in women twins: heritability and relationship to apolipoprotein(a) phenotypes. Am J Hum Genet. 1992;51: Luke MM, Kane JP, Liu DM, et al. A polymorphism in the protease-like domain of apolipoprotein(a) is associated with severe coronary artery disease. Arterioscler Thromb Vasc Biol. 2007;27: Kamstrup PR, Tybjaerg-Hansen A, Steffensen R, Nordestgaard BG. Pentanucleotide repeat polymorphism, lipoprotein(a) levels, and risk of ischemic heart disease. J Clin Endocrinol Metab. 2008;93: Berglund L, Ramakrishnan R. Lipoprotein(a): an elusive cardiovascular risk factor. Arterioscler Thromb Vasc Biol. 2004;24: Concentration of circulating lipoprotein(a) is an independent heritable factor associated with risk of coronary artery disease. The rs single nucleotide polymorphism that results in a null allele is not included in any of the widely used commercial high-throughput single nucleotide polymorphism arrays and has not been previously examined in large-scale genetic epidemiological studies. Here, we show the association of this functional single nucleotide polymorphisms with reduced lipoprotein(a) levels and risk of coronary artery disease and highlight its importance for future genetic and epidemiological studies.